chapter5(link) ece358slides

7/31/2019 Chapter5(Link) ECE358Slides

1/182

Data Link Layer 5-1

Chapter 5

Link Layer and LANs

A note on the use of these ppt slides:Were making these slides freely available to all (faculty, students, readers).

Theyre in PowerPoint form so you can add, modify, and delete slides

(including this one) and slide content to suit your needs. They obviously

represent a lotof work on our part. In return for use, we only ask thefollowing:

If you use these slides (e.g., in a class) in substantially unaltered form, thatyou mention their source (after all, wed like people to use our book!)

If you post any slides in substantially unaltered form on a www site, that

you note that they are adapted from (or perhaps identical to) our slides, and

note our copyright of this material.

Thanks and enjoy! JFK/KWR

All material copyright 1996-2010

J.F Kurose and K.W. Ross, All Rights Reserved

Computer Networking:A Top Down Approach5th edition.Jim Kurose, Keith RossAddison-Wesley, April2009.


2/182

Data Link Layer 5-2

Chapter 5: The Data Link Layer

Our goals: understand principles behind data link layerservices:

errordetection and correction ( by the receiver)

link access by sharing a broadcast channel: multiple access Instruct the hardware (PHY layer) whento transmit ( MAC protocols)

link layer addressing

reliable data transfer, flow control

implementation of various link layer technologies


3/182

Data Link Layer 5-3

Link Layer

5.1 Introduction andservices

5.2 Error detection andcorrection

5.3Multiple access protocols

New: WiFi

5.4 Link-layer Addressing

5.5 Ethernet

5.6 Link-layer switches

5.7 PPP (Point-to-Point Protocol)

5.8 Link virtualization: MPLS(Multi-protocol Label Switching)

5.9 A day in the life of a webrequest (dropped: NOTmeaningful in bottom/upapproach to understandingnetworking)

3.4 New: Reliable datatransfer + flow control


4/182

Data Link Layer 5-4

Link Layer: Introduction

Terminology: hosts and routers are nodes communication channels that

connect adjacent nodes are links wired links

wireless links LANs

layer-2 packet is a frame,encapsulates datagram (from netlayer)

data-link layer has responsibility oftransferring datagram from one node

tophysically adjacentnode over a link


5/182

Data Link Layer 5-5

Link layer: context

datagram transferred bydifferent link protocolsover different links: e.g., Ethernet on first link,

frame relay on intermediate

links, 802.11 on last link each link protocol provides

different services e.g., may or may not provide

reliable data transfer overlink

transportation analogy trip from Princeton to

Lausanne

limo: Princeton to JFK

plane: JFK to Geneva

train: Geneva to Lausanne tourist = datagram

transport segment =communication link

transportation mode =link layer protocol

travel agent = routingalgorithm


6/182

Data Link Layer 5-6

Link Layer Services

framing, link access: encapsulate datagram into frame, adding header, trailer

channel access if shared medium (a.k.a. broadcast medium)

MAC addresses used in frame headers to identify source,

dest different from IP address!

reliable delivery between adjacent nodes we learned how to do this already (ch. 3)! (Will do it in this ch.)

seldom used on low bit-error link (fiber, some twisted pair) wireless links: high error rates

Q.: Why both link-level and end-to-end reliability?


7/182Data Link Layer 5-7

Link Layer Services (more)

flow control: pacing between (adjacent) sending and receiving nodes

error detection: errors caused by signal attenuation and noise

receiver detects presence of errors error correction:

receiver identifies and correctsbit error(s) withoutresorting to retransmission

receiver signals sender for retransmission

half-duplex and full-duplex with half duplex, nodes at both ends of link can transmit,

but not at the same time



Where is the link layer implemented?

in each and every host link layer implemented in

adaptor (aka networkinterface cardNIC) Ethernet card, PCMCI

card, 802.11 card implements link, physical

layer

attaches into hosts

system buses combination of

hardware, software,firmware

controller

physical

transmission

cpu memory

hostbus(e.g., PCI)

network adaptercard

host schematic

application

transport

networklink

link

physical



Adaptors Communicating

sending side: encapsulates datagram in

frame

adds error checking bits,rdt, flow control, etc.

receiving side looks for errors, rdt, flow

control, etc

extracts datagram, passesto upper layer at receivingside

controller controller

sending host receiving host

datagram datagram

datagram

frame



Link Layer



5.3Multiple accessprotocols

5.4 Link-layerAddressing

5.5 Ethernet


5.7 PPP

5.8 Link virtualization:

MPLS5.9 A day in the life of aweb request

3.4 Reliable data transfer


11/182


12/182

Data Link Layer 5-12

Parity Checking

Single Bit Parity:Detect single bit errors

Two Dimensional Bit Parity:Detect and correctsingle bit errors

0 0


13/182


Internet checksum (review; detailsin Transport Layer . Just wait.)

Sender: treat segment contents

as sequence of 16-bitintegers

checksum: addition (1scomplement sum) of

segment contents sender puts checksum

value into UDP checksumfield

Receiver:

compute checksum ofreceived segment

check if computed checksumequals checksum field value:

NO - error detected

YES - no error detected.But maybe errorsnonetheless?

Goal:detect errors (e.g., flipped bits) in transmittedpacket (note: used at transport layer only)


14/182


15/182


CRC Example: D = 101110, r = 3, G = 1001

Want:D.2rXOR R = nG

equivalently:

D.2r = nG XOR R

equivalently:if we divide D.2r by G,we get remainder R

R = remainder[ ]D.2r

G


16/182


1 0 1 1 1 0 0 0 01 0 0 1------------------0 0 1 0 10 0 0 0

1 00 1------------------0 0 1 0 10 0 0 0

1 0 01------------------0 0 0 0 11 0 0 0

1 00 1------------------0 0 0 0 11 0 0 0

10 0 1------------------0 0 0 0 01 0 1 0

1 0 0 1------------------0 0 0 0 000 1 1

Process: Calculation of CRCIf the input bit above the leftmost divisor bit is 1,

the divisor is XORed into the input.Else (the input bit above the leftmost divisor bit is 0) do nothing.

The divisor is then shiftedone bit to the right (

)The process is repeated untilthe divisor reaches the right-hand end of the input row.

Input: D.2r

Divisor: G

Align input and divisor on MSB

R


17/182


Link Layer





5.5 Ethernet


5.7 PPP


MPLS5.9 A day in the life of aweb request


18/182


Multiple Access Links and Protocols

Two types of links: point-to-point

PPP for dial-up access

point-to-point link between Ethernet switch and host

broadcast (shared wire or medium) old-fashioned Ethernet upstream HFC

802.11 wireless LAN

shared wire (e.g.,cabled Ethernet)

shared RF(e.g., 802.11 WiFi)

shared RF(satellite)

humans at acocktail party

(shared air, acoustical)


19/182


Multiple Access protocols

single shared broadcast channel

two or more simultaneous transmissions by nodes:interference collision if node receives two or more signals at the same time

multiple access protocol

distributed algorithm that determines how nodesshare channel, i.e., determine when a node can transmit

communication about channel sharing must use channelitself! no separate control channel for coordination


20/182

20

PacketCollision

PHY

Link/MACdata

PHY

Link/MACdata

PHY

Link/MAC

+

Rx

Tx

Tx


21/182


Ideal Multiple Access Protocol

Broadcast channel of rate R bps

1. when one node wants to transmit, it can send at rate R.

2. when M nodes want to transmit, each can send ataverage rate R/M

3. fully decentralized:

no special node to coordinate transmissions no synchronization of clocks, slots

4. simple


22/182


MAC Protocols: a taxonomy

Three broad classes: Channel Partitioning (Commonly done in cellular networks)

divide channel into smaller pieces (time slots, frequency, code)

allocate piece to node for exclusive use

Random Access channel not divided, allow collisions

recover from collisions

Taking turns

nodes take turns


23/182


Channel Partitioning MAC protocols: TDMA

TDMA: time division multiple access access to channel in "rounds"

each station gets fixed length slot (length = pkttrans time) in each round

unused slots go idle example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6

idle

1 3 4 1 3 4

6-slotframe


24/182


Channel Partitioning MAC protocols: FDMA

FDMA: frequency division multiple access channel spectrum divided into frequency bands

each station assigned fixed frequency band

unused transmission time in frequency bands go idle

example: 6-station LAN, 1,3,4 have pkt, frequencybands 2,5,6 idle

frequencyb

ands

FDM cable


25/182


Random Access Protocols

When node has packet to send transmit at full channel data rate R.

no a prioricoordination among nodes (partial exception in WiFi)

two or more transmitting nodes collision,

random access MAC protocol specifies: how to detect collisions

how to recover from collisions (e.g., via delayed retransmissions)

Examples of random access MAC protocols: ALOHA and slotted ALOHA

CSMA/CD, CSMA/CA CSMA: Carrier Sense Multiple Access CD: Collision Detection

CA: Collision Avoidance


26/182

26

Aloha Protocol

Developed in the 1970s at U of Hawaii

To interconnect terminals with mainframes

LAN/ WLAN: Possible, but not used

GSM: Cell phones use this protocol to request achannel from the base stations

Two types Pure Aloha (Continuous time)

Slotted Aloha


27/182

27

Pure AlohaStart:i = 0

Start atimerT = 2*Tp +

ACKreceivedCancel timer

Success

Timeouti++

TransmitFrame

i > Kmax

R =Random(0,2i-1)

WaitTB= R*Tp

Error

Yes

No

Exponential backoff


28/182


Pure (unslotted) ALOHA

unslotted Aloha: simpler, no synchronization when frame first arrives

transmit immediately

collision probability increases:

frame sent at t0 collides with other frames sent in [t0-1,t0+1]


29/182


Pure Aloha efficiency

P(success by given node) = P(node transmits) .P(no other node transmits in [t0-1,t0] .P(no other node transmits in [t0, t0+1]

= p . (1-p)N-1 . (1-p)N-1

= p . (1-p)2(N-1)

choosing optimum p and then letting n -> infty ...

= 1/(2e) = .18

even worsethan slotted Aloha!


30/182


Slotted ALOHA

Assumptions: all frames same size time divided into equal

size slots (time to

transmit 1 frame) nodes start to transmit

only at slot beginning nodes are synchronized

if 2 or more nodestransmit in a slot, allnodes detect collision

Operation: when node obtains fresh

frame, transmits in nextslot

if no collision:node cansend new frame in nextslot

if collision:noderetransmits frame ineach subsequent slotwith prob. p untilsuccess


31/182


Slotted ALOHA

Pros

single active node cancontinuously transmitat full rate of channel

highly decentralized:only slots in nodesneed to be in sync

simple

Cons collisions, wasting slots idle slots nodes may be able to

detect collision in lessthan time to transmitpacket

clock synchronization


32/182


Slotted Aloha efficiency

suppose:N nodes with manyframes to send, eachtransmits in slot withprobabilityp

prob that given node has

success in a slot = p(1-p)N-1

prob that anynode has asuccess = Np(1-p)N-1

max efficiency: find p* thatmaximizesNp(1-p)N-1

for many nodes, take limit ofNp*(1-p*)N-1 as N goes to

infinity, gives:Max efficiency = 1/e = .37

Efficiency : long-run

fraction of successful slots(many nodes, all with manyframes to send)

At best:channelused for usefultransmissions 37% of time! !


33/182


CSMA (Carrier Sense Multiple Access)

CSMA: listen before you transmit:

If channel sensed idle: transmit entire frame

If channel sensed busy, defer transmission

human analogy: dont interrupt others!


34/182

34

CSMA/CDConcepts of Carrier Sense and Collision Detection

Tx Rx

Data

Collision?Carrier?

Medium

MAC/PHY

Sense voltage (V) on medium:V > Vth1carrier is presentV > Vth2 collision

Time

Vth2

Vth1

V


35/182


CSMA collisions

collisions canstill occur:propagation delay meanstwo nodes may not heareach others transmission

collision:entire packet transmissiontime wasted

note:role of distance & propagationdelay in determining collisionprobability


36/182


CSMA/CD (Collision Detection)

CSMA/CD: carrier sensing, deferral as in CSMA collisions detectedwithin short time

colliding transmissions aborted, reducing channelwastage

collision detection: easy in wired LANs: measure signal strengths

difficult in wireless LANs


37/182

37

CSMA/CD Start:i = 0

Success

Sendjamming signal+Abort

i++

Sensemedium

i > Kmax

R = Random(0,2i-1)

WaitTB= R*Tp

Error

Yes

No

Busy WaitY

N

No collisionCollision

TransmitframeWHILEdetectingcollision

Exponential backoff


38/182

38

CSMA/CD

Medium sensing is done for 96 bit-times.

Jamming signal length is 48 bits. Jamming signal createsenough energy on the medium for collision detection.

Tp is equated with 512 bit-times.

i saturates at 10.


39/182


CSMA/CD collision detection


40/182


Taking Turns MAC protocols

channel partitioning MAC protocols:

share channel efficientlyand fairlyat high load

inefficient at low load: delay in channel access,1/N bandwidth allocated even if only 1 activenode!

random access MAC protocols

efficient at low load: single node can fullyutilize channel

high load: collision overhead

taking turns protocols

look for best of both worlds!


41/182



Polling: master node

invites slave nodesto transmit in turn

typically used withdumb slave devices

concerns: polling overhead

latency single point of

failure (master)

master

slaves

poll

data

data


42/182



Token passing: control token passed

from one node to nextsequentially.

token message concerns:

token overhead

latency

single point of failure(token)

T

data

(nothingto send)

T


43/182

43

Wireless LAN

IEEE 802.11/a/b/g


44/182

44

WLAN View

Access

Point

C

C

C: Computer, AP: Access Point

Radio range of the AP

IEEE 802.11 protocol

Basic Service Set (BSS): BSSID = MAC address of APIndependent BSS (IBSS)= BSS - APExtended Service Set (ESS): A collection of BSS

connected by a Distribution System


45/182

45

IEEE 802.11/a/b/g/n Family

IEEE Technique Frequency

Band

Rate

(Mbps)

802.11 DSSS

FHSS

2.4 GHz

2.4 GHz

1 and 2

1 and 2

802.11a OFDM 5 GHz 6--54802.11b DSSS 2.4 GHz 5.5 and 11

802.11g

802.11n

802.11ac

(Draft/Nov. 2011)

OFDM

OFDM

OFDM

2.4 GHz

2.4/5 GHz

5 GHz

22 and 54

72 and 150

6.9 Gbps


46/182

46

Different Modes of Operation of MAC in IEEE802.11

Modes of IEEE802.11 MAC

Point Coordination

Function(PCF)modeDistributed Coordination

Function(DCF)mode

With Hand-shake Without Hand-shake


47/182

47

PCF Mode: Optional

The AP Operates as the central controller for all nodes within its range. Decideswho transmits and when.

Can follow a round-robin policy to allocate slots.

Note:There is no contention for medium access.

This mode Can support real-time traffic due to periodic scheduling.

Leads to waste of bandwidth if a scheduled node has no traffic.

Is optional


48/182

48

DCF Mode: Mandatory

An AP Need not be used.

Computers can directly communicate among themselves


49/182

49

Alternating use of PCF and DCF

The WLAN operates In the PCF mode for T1 seconds

Bandwidth guarantee for some nodes

In the DCF mode for T2 seconds Nodes with additional traffic can contend for a share of the bandwidth

PCF DCF PCF DCF PCF DCF

TimeT1 T2 T3 T4


50/182

50

DCF with hand-shake A sender obtains permission from the receiver

before transmitting a data frame. called hand-shake

Hand-shake mechanism Sender transmits a Request To Send (RTS) frame

Receiver gives permission by sending back a Clear To Send (CTS)frame

Used to increase the probability of successful Tx when

Packet length is long. ( dotRTSThreshold holds the value.)

Incurs additional cost

loss of some bandwidth due to hand-shake


51/182

51

DCF with and without hand-shake

The two modes are not mutually exclusive. A node decides what mode to use on a frame/frame basis. The MAC management database contains a variable

dotRTSThreshold: integer in bytes

Length of a data frame >= dotRTSThreshold

Use hand-shake

Length of a data frame < dotRTSThreshold

Do not use hand-shake

PCF PCF

DCF

Hand-shake

No hand-shake

PCF

DCF

Mode of operation of the same node}


52/182

52

DCF withouthand-shake A sender does not obtain permission from the

receiver before transmitting a data frame.

RTS/CTS mechanism is NOTused. There is no prior coordination between sender and receiver

A sender transmits a frame when some medium sensing conditions are satisfied.

To follow

When data frames are short Use this to save bandwidth


53/182

53

Problems in WLAN

Hidden Terminal Problem Exposed Terminal Problem

Inability to detect collision (at the receiver)

Assumption All nodes have identical radio ranges

how far away their signals can be received

Note The assumption is not the cause of the problems.

Without this assumption, the problems become worse.


54/182

Exp sed Terminal Pr blem


55/182

55

Exposed Terminal Problem

A B C

Tx

Problem- A istransmittinga frame to D.- B knowsthat someone is transmitting.

- If B transmits a frame to C, no problem. However, Bdoes not transmitbecause it is unaware of Ds location.

The above problem is due to B beingexposedto As Tx.

D


56/182

56

No collision detection

Fact:Collision at receiver is more serious than

collision at transmitter. In a wired LAN Collision is indirectlydetected after some delay by

the sender. In a WLAN, collision detection by sender is not

possible. That is why we have the hidden terminal problem.

Collision is avoided (CA), rather than detected

WLAN MAC: CSMA/CA


57/182

57

WLAN MAC: CSMA/CA (RTS: Request to Send; CTS: Clear to Send)

In CSMA/CA, collision is avoided using PHY-level carrier sensing: Done in receiver hardware Virtual carrier sensing: Done by examining frame header

Frame headers of RTS, CTS, DATA A durationfield in frame headers indicates

for how long the sender of the frame may use the medium.

A Network Allocation Vector is managed using durationfields Each node has its own NAV (essentially an integer)

NAV > 0: A node had announced its intention to use the medium now.

NAV = 0:Nobody had wanted to use the medium now.

Transmit condition: When medium is idle(Carrier is absent)AND(NAV = 0)


58/182

58

NAV Update Mechanism

Each node has its own NAV. NAV is the length of time for which the medium is likely

to remain busy.

Procedure to update NAV Initially: NAV = 0. With each passing s (micro second)

NAV = NAV 1 Stop decrementing ifNAV = 0.

NAV is updated using the durationfield in a received

frame NAV = Max(NAV, duration)


59/182

59

RTS and CTS Frames

Frame format

FrameControl

Duration RA TA FCSRTS

FrameControl

Duration RA FCSCTS/ACK

2 2 6 6 4 bytes

2 2 6 4 bytes

FCS: Frame Check Sequence CRCRA: Receiver AddressTA: Transmitter Address


60/182

60

DATA Frame

Frame format

FrameControl

Duration/ID

A1 A2 A3 A4Seq.

ControlFCSFrame Body

RA TA

TA:Physically transmitting the frame.

TA RA

Ti i I l


61/182

61

Timing Intervals The IEEE 802.11 MAC defines 4 timing intervals 2 at the PHY level

SIFS: Short Inter-Frame Space (10 micro-sec)

aSlot (20 micro-sec)

2 at the MAC level PIFS: Priority (in PCF) IFS (SIFS + aSlot)

DIFS: Distributed IFS (PIFS + aSlot)

Note:aSlot is chosen s.t. a station is capable of determining if another

station initiated a Tx at the beginning of the previous slot.

H d h k i RTS/CTS


62/182

62

Hand-shake using RTS/CTS

RTS

CTSSIFS

SIFSDATA

SIFS ACK

Value of duration in RTS

Value of duration in CTS

A

B

C(neighbor of A)

D

(neighbor of B)

Value of NAV of C

Value of NAV of D

Time

DIFS

data

DCF with Hand-shake: Tx


63/182

63

DCF with Hand shake: TxF: a new data frame to be transmittedi = 0, CW = CWmin

NAV =0?

Send RTSStart a timer

Cancel timer

Send DATA (F)Start a timer

i = i+1

CW = CWmin*2i

(At some point, CW saturates at CWmax.)

Timeout

Timeout

RandomBackoff

No

Yes

CTS is received

ACK is received

End ofbackoff

Yes

No

Medium idlefor DIFS?

B k ff M h i


64/182

64

Backoff Mechanism

Initialize a counter: Backoff Time Counter (BTC)

BTC = Random(0,CW-1) The time unit of BTC is aSlotTime

aSlotTime: propagation + transceiver switching time

As time passes, BTC is decremented as follows BTC = BTC -1 if medium is idle for aSlotTime

Pause decrementing BTCif medium is busy

Resume decrementingBTCif medium is idle for DIFS.

Subsequent decrementing is done for every aSlotTime ofidleness of the medium.

B k ff M h i


65/182

65

Backoff Mechanism

Ch. busy dueto node A

Ch. Busy due tonode C

DIFS DIFSX X X X

BTC = 5

5 4 3 3 2

3

X = aSlotTime

B 1 0

Time

Bis executingbackoff


66/182

66

DCF with Hand-shake: Rx

Receive RTS

NAV = 0?

Send CTS

No

Yes

No

Yes

Receive DATA frame

Send data to upperlayer

Send ACK

Note:The above two fragments of flow-charts can

be easily merged.

No

Yes

Medium idle forSIFS?

Medium idle forSIFS?

DCF Mode without Hand shake


67/182

67

DCF Mode without Hand-shake

A special case of DCF with hand-shake

RTS/CTS frames are not exchanged. The idea of NAV is still used in this mode All nodes process the received RTS/CTS of others

NOTE: A node may broadcast a DATA frame to all Done in DCF without hand-shake

Receivers do not send back an ACK.

Withouthandshake

RTS/CTS This must processreceived RTS/CTS

PCF Mode: AP becomes the controller


68/182

68

AP alternates between PCF and DCF modes

AP operates as the controller as follows AP senses the medium at the start of a CF (Contention-

Free) period for a PIFS (Priority IFS) interval. SIFS < PIFS < DIFS

You know that

PIFS = SIFS + aSlotTime

DIFS = SIFS + 2*aSlotTime

If the medium is idle for PIFS, transmit a beacon frame Beacon contains a CFPMaxDuration field

Nodes receiving a beacon update their NAV to CFPMaxDuration

These nodes perceive the medium to be busy for CFPMaxDuration

PCF Mode of Operation (Contd )


69/182

69

PCF Mode of Operation (Contd.)

After transmitting a beacon, AP waits for SIFS beforetransmitting one of the following

DATA frame

CF Poll frame

DATA+CF Poll frame ACK frame

CF End frame


70/182

70


CF Poll frame

AP APUser 1 User 2 User 1

CF Poll CF Poll

DATA DATA

ACK ACK

The polled usersends data toanother user.

The polled user sendsdata to the AP.

SIFS

SIFS

SIFS

SIFS

Time


71/182

71


DATA + CF Poll frame

AP User 1 User 2DATA+CFPoll

DATA

ACK

The polled user receives data from the AP andsends data to another user.

SIFS

SIFS

ACK

Note 1:If AP doesnotreceive an ACK, itretransmits data after PIFS.

Note 2:If User 1 does not receive ACK, it doesnot retransmitdata.



72/182

72


DATA frame

Contains user data from AP to a specific station. The receiver sends back an ACK after SIFS interval.

AP does not receive an ACK Retransmit the DATA after a PIFS interval

AP can broadcast a DATA frame These are not ACKed.



73/182

73


CF Poll frame

AP grants permission to another node to transmit DATAto the AP or to a third node.

Receiver of DATA frame sends an ACK to the sender.

If the polled node has no data to send, it sends a null

DATA frame. If the polled station does not receive an ACK, it can not

retransmit its data frame until it is polled again.


74/182

A d j i i WLAN ith AP


75/182

75

A node joining a WLAN with an AP

(No need for such a procedure in a wired LAN) Two ways for a node to join a WLAN Passive scanning

Scan a channel for a Beacon frame

If a Beacon frame is received NegotiateAuthentication andAssociation processes

Active Scanning Transmit a Probe frame

If a Probe Response is received Negotiate Authentication and Authorization processes

Summary of MAC protocols


76/182


Summary of MAC protocols

channel partitioningby time, frequency or code Time Division, Frequency Division, Code Division

random access(dynamic), ALOHA, S-ALOHA, CSMA/CD

carrier sensing: easy in some technologies (wire), hard in wireless

CSMA/CD used in Ethernet

CSMA/CA used in 802.11

taking turns polling from central site, token passing

Bluetooth, FDDI (Fibre Dist. Data Interface), IBM Token Ring

Bluetoothpiconet

master

Up to 7 slaves

Li k L


77/182


Link Layer




5.4 Link-LayerAddressing

5.5 Ethernet

5.6 Link-layer switches5.7 PPP

5.8 Link virtualization:MPLS

5.9 A day in the life of aweb request


MAC Add d ARP


78/182


MAC Addresses and ARP (Address Resolution Protocol)

32-bit IP address: (for comparison only .) network-layeraddress

used to get datagram to destination IP subnet

MAC (or LAN or physical or Ethernet) address: function:get frame from one interface to another

physically-connected interface(same network)

48 bit MAC address (for most LANs) burned in NIC ROM, also sometimes software settable


79/182

LAN Add d ARP


80/182


LAN Addresses and ARP

Each adapter on LAN has unique LAN address

Broadcast address =FF-FF-FF-FF-FF-FF

= adapter

1A-2F-BB-76-09-AD

58-23-D7-FA-20-B0

0C-C4-11-6F-E3-98

71-65-F7-2B-08-53

LAN(wired orwireless)

LAN Address (more)


81/182


LAN Address (more)

MAC address allocationadministered by IEEE

manufacturer buys portion of MAC address space(to assure uniqueness)

MAC flat address portability can move LAN card from one LAN to another

(IP hierarchical address NOT portable) address depends on IP subnet to which node is attached

ARP: Address Resolution Protocol


82/182


ARP: Address Resolution Protocol

Each IP node (host, router)

on LAN has ARP table

ARP table: IP/MAC addr.mappings for some LAN

nodes< IP address; MAC address; TTL>

TTL (Time To Live): timeafter which address mappingwill be forgotten (typically 20min)

Question:

how to determine Bs MAC addressknowing Bs IP address?

1A-2F-BB-76-09-AD

58-23-D7-FA-20-B0

0C-C4-11-6F-E3-98

71-65-F7-2B-08-53

LAN

137.196.7.23

137.196.7.78

137.196.7.14

137.196.7.88

ARP protocol: Same LAN (network)


83/182


ARP protocol Same LAN (network)

A wants to send datagram

to B, and Bs MAC addressnotin As ARP table.

A broadcastsARP querypacket, containing B's IPaddress

dest MAC address =FF-FF-FF-FF-FF-FF

all machines on LANreceive ARP query

B receives ARP query,replies to A with its (B's)MAC address frame sent to As MAC

address (unicast)

A caches (saves) IP-to-MACaddress pair in its ARP tableuntil information becomes old(times out)

soft state: information

that times out (goes away)unless refreshed

ARP is plug-and-play: nodes create their ARP

tables without interventionfrom net administrator

Addressing: routing to another LAN


84/182


walkthrough: send datagram from A to B via R.

focus on addressing - at both IP (datagram) and MAC layer (frame) assume A knows Bs IP address

assume A knows Bs MAC address (how?)

assume A knows IP address of first hop router, R (how?)

assume A knows MAC address of first hop router interface (how?)

Addressing routing to another LAN

R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55

A

222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

B



85/182



R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55

A

IPEth

Phy

IP src: 111.111.111.111

IP dest: 222.222.222.222

A creates IP datagram with IP source A, destination B

A creates link-layer frame with R's MAC address as dest,frame contains A-to-B IP datagram

MAC src: 74-29-9C-E8-FF-55

MAC dest: E6-E9-00-17-BB-4B

222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

B



86/182



R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55

A

IPEth

Phy

frame sent from A to R

IP src: 111.111.111.111

IP dest: 222.222.222.222

MAC src: 74-29-9C-E8-FF-55

MAC dest: E6-E9-00-17-BB-4B

IPEth

Phy

frame received at R, datagram removed, passed up to IP

222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

B



87/182



R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55 222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

BA

IP src: 111.111.111.111

IP dest: 222.222.222.222

R forwards datagram with IP source A, destination B

R creates link-layer frame with B's MAC address as dest,frame contains A-to-B IP datagram

MAC src: 1A-23-F9-CD-06-9B

MAC dest: 49-BD-D2-C7-56-2A

IPEth

Phy

IP

EthPhy



88/182


Address ng rout ng to another LAN

R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55 222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

BA



IP src: 111.111.111.111

IP dest: 222.222.222.222



IPEth

Phy

IP

EthPhy



89/182


ng u ng n L N

R

1A-23-F9-CD-06-9B222.222.222.220

111.111.111.110E6-E9-00-17-BB-4BCC-49-DE-D0-AB-7D

111.111.111.112

111.111.111.111

74-29-9C-E8-FF-55 222.222.222.22249-BD-D2-C7-56-2A

222.222.222.22188-B2-2F-54-1A-0F

BA



IP src: 111.111.111.111

IP dest: 222.222.222.222



IP

EthPhy

Link Layer


90/182


Link Layer





5.5 Ethernet




Ethernet


91/182


Ethernet

dominant wired LAN technology:

cheap $20 for NIC

first widely used LAN technology

simpler, cheaper than token LANs and ATM

kept up with speed race: 10 Mbps 10 Gbps

Metcalfes Ethernet

sketch

Star topology


92/182


p gy

bus topology popular through mid 90s

all nodes in same collision domain (can collide with eachother)

today:star topology prevails active switchin center each spoke runs a (separate) Ethernet protocol (nodes

do not collide with each other)

switch

bus: coaxial cable star

Ethernet Frame Structure


93/182


Ethernet Frame Structure

Sending adapter encapsulates IP datagram (or othernetwork layer protocol packet) in Ethernet frame

Preamble:

7 bytes with pattern 10101010 followed by one bytewith pattern 10101011

used to synchronize receiver, sender clock rates

Ethernet Frame Structure (more)


94/182


Ethernet Frame Structure (more)

Addresses: 6 bytes if adapter receives frame with matching destination

address, or with broadcast address (e.g. ARP packet), itpasses data in frame to network layer protocol

otherwise, adapter discards frame

Type: indicates higher layer protocol (mostly IPbut others possible, e.g., Novell IPX, AppleTalk)

CRC: checked at receiver, if error is detected,frame is dropped

Ethernet: Unreliable connectionless


95/182


Ethernet: Unreliable, connectionless

connectionless: No handshaking between sending andreceiving NICs

unreliable:receiving NIC doesnt send acks or nacksto sending NIC

stream of datagrams passed to network layer can have gaps(missing datagrams)

gaps will be filled if app is using TCP

otherwise, app will see gaps

Ethernets MAC protocol: unslotted CSMA/CD

Ethernet CSMA/CD algorithm


96/182


g

1. NIC receives datagramfrom network layer,creates frame

2. If NIC senses channel idle,starts frame transmission.

If NIC senses channelbusy, waits until channelidle, then transmits

3. If NIC transmits entire

frame without detectinganother transmission, NICis done with frame !

4. If NIC detects anothertransmission whiletransmitting, aborts andsendsjam signal

5. After aborting, NIC

enters exponentialbackoff: after mthcollision, NIC chooses Katrandom from

{0,1,2,,2m

-1}. NIC waitsK512 bit times, returns toStep 2

Ethernets CSMA/CD (more)


97/182


Ethernet s CSMA/CD (more)

Jam Signal: make sure allother transmitters areaware of collision; 48 bits

Bit time: .1 microsec for 10Mbps Ethernet ;for K=1023, wait time is

about 50 msec

Exponential Backoff:

Goal: adapt retransmissionattempts to estimatedcurrent load

heavy load: random waitwill be longer

first collision: choose K from{0,1}; delay is K 512 bittransmission times

after second collision: choose

K from {0,1,2,3} after ten collisions, choose K

from {0,1,2,3,4,,1023}

See/interact with Java

applet on AWL Web site:highly recommended !

CSMA/CD efficiency


98/182


CSMA/CD efficiency

Tprop = max prop delay between 2 nodes in LAN ttrans = time to transmit max-size frame

efficiency goes to 1 as tprop goes to 0

as ttrans goes to infinity better performance than ALOHA: and simple,

cheap, decentralized!

transprop/ttefficiency 51

1

802.3 Ethernet Standards: Link & Physical Layers


99/182


y y

manydifferent Ethernet standards common MAC protocol and frame format different speeds: 2 Mbps, 10 Mbps, 100 Mbps,

1Gbps, 10G bps different physical layer media: fiber, cable

application

transport

networklink

physical

MAC protocol

and frame format

100BASE-TX

100BASE-T4

100BASE-FX100BASE-T2

100BASE-SX 100BASE-BX

fiber physical layercopper (twisterpair) physical layer

Manchester encoding


100/182

Data Link Layer5-100

Manchester encoding

used in 10BaseT

each bit has a transition

allows clocks in sending and receiving nodes tosynchronize to each other no need for a centralized, global clock among nodes!

Hey, this is physical-layer stuff!

Link Layer


101/182


Link Layer



5.3 Multiple accessprotocols


5.5 Ethernet

5.6 Link-layer switches,LANs, VLANs

5.7 PPP


MPLS5.9 A day in the life of a

web request

Hubs


102/182


physical-layer (dumb) repeaters:

bits coming in on one link go out allother links at samerate

all nodes connected to hub can collide with one another

no frame buffering

no CSMA/CD at hub: host NICs detect collisions

twisted pair

hub

Switch


103/182


link-layer device: smarter than hubs, take

activerole store, forward Ethernet frames

examine incoming frames MAC address,selectively forward frame to one-or-more

outgoing links when frame is to be forwarded onsegment, uses CSMA/CD to access segment

transparent hosts are unaware of presence of switches

plug-and-play, self-learning switches do not need to be configured

Switch: allows multiplesimultaneous


104/182


ptransmissions

hosts have dedicated,direct connection to switch

switches buffer packets Ethernet protocol used on

eachincoming link, but nocollisions; full duplex each link is its own collision

domain

switching:A-to-A and B-to-B simultaneously,without collisions not possible with dumb hub

A

A

B

B

C

C

switch with six interfaces(1,2,3,4,5,6)

1 23

45

6

Switch Table


105/182


Switch Table

Q:how does switch know thatA reachable via interface 4,B reachable via interface 5?

A: each switch has a switch

table, each entry: (MAC address of host, interface

to reach host, time stamp)

looks like a routing table!

Q:how are entries created,maintained in switch table? something like a routing

protocol?

A

A

B

B

C

C

switch with six interfaces(1,2,3,4,5,6)

1 23

45

6

Switch: self-learning

Source: AD t A


106/182


Switch: self learning

switchlearnswhich hostscan be reached throughwhich interfaces when frame received,

switch learns location ofsender: incoming LANsegment

records sender/locationpair in switch table

A

A

B

B

C

C

1 23

45

6

A A

Dest: A

MAC addr interface TTL

Switch table(initially empty)

A 1 60

Switch: frame filtering/forwarding


107/182


When frame received:

1. record link associated with sending host

2. index switch table using MAC dest address

3. ifentry found for destination

then {ifdest on segment from which frame arrived

then drop the frame

else forward the frame on interface indicated

}elseflood forward on all but the interface

on which the frame arrived

Self-learning,

Source: AD st: A


108/182


f g,forwarding:

example

A

A

B

B

C

C

1 23

45

6

A A

Dest: A

MAC addr interface TTL

Switch table(initially empty)

A 1 60

A AA AA AA AA A

frame destinationunknown:flood

A A

destination Alocation known:

A 4 60

selective send

Interconnecting switches


109/182


switches can be connected together

A

B

Q:sending from A to G - how does S1 know toforward frame destined to F via S4 and S3?

A:self learning! (works exactly the same as insingle-switch case!)

S1

C DE

FS2

S4

S3

HI

G

Self-learning multi-switch example


110/182


Suppose C sends frame to I, I responds to C

Q:show switch tables and packet forwarding in S1,S2, S3, S4

A

B

S1

C DE

FS2

S4

S3

HI

G

1

2

Institutional network


111/182


Institutional network

to externalnetwork

router

IP subnet

mail server

web server

Switches vs. Routers


112/182


both store-and-

forward devices routers: network-layer

devices (examinenetwork-layer headers)

switches are link-layerdevices (examine link-

layer headers) routers maintain

routing tables,implement routingalgorithms

switches maintainswitch tables,implement filtering,learning algorithms

applicationtransport

networklinkphysical

networklink

physical

linkphysical

switch

datagram

application

transportnetwork

linkphysical

frame

frame

frame

datagram

VLANs: motivation


113/182


VLANs motivation

What happens if: CS user moves office to EE,

but wants connect to CSswitch?

single broadcast domain: all layer-2 broadcast

traffic (ARP, DHCP)crosses entire LAN(security/privacy,efficiency issues)

each lowest level switch hasonly few ports in use

ComputerScience Electrical

Engineering

ComputerEngineering

Whats wrong with this picture?

VLANs Port-based VLAN: switch ports grouped(by switch management software) so


114/182


VLANs (by switch management software) sothat singlephysical switch

Switch(es) supportingVLAN capabilities canbe configured todefine multiple virtualLANS over singlephysical LANinfrastructure.

Virtual LocalArea Network

1

8

9

16102

7

Electrical Engineering

(VLAN ports 1-8)

Computer Science

(VLAN ports 9-15)

15


(VLAN ports 1-8)

1

82

7 9

1610

15

Computer Science

(VLAN ports 9-16)

operates as multiplevirtual switches

Port-based VLAN


115/182


Port based VLAN

1

8

9

16102

7


(VLAN ports 1-8)

Computer Science

(VLAN ports 9-15)

15

traffic isolation:frames

to/from ports 1-8 canonlyreach ports 1-8 can also define VLAN based on

MAC addresses of endpoints,rather than switch port

dynamic membership:ports can be dynamicallyassigned among VLANs

router

forwarding between VLANS:

done via routing (just as withseparate switches) in practice vendors sell combined

switches plus routers

VLANS spanning multiple switches


116/182


L p g m p

trunk port:carries frames between VLANS definedover multiple physical switches

frames forwarded within VLAN between switches cant bevanilla 802.1 frames (must carry VLAN ID info)

802.1q protocol adds/removed additional header fields forframes forwarded between trunk ports

1

8

9

102

7


(VLAN ports 1-8)

Computer Science

(VLAN ports 9-15)

15

2

73

Ports 2,3,5 belong to EE VLAN

Ports 4,6,7,8 belong to CS VLAN

5

4 6 816

1

802.1Q VLAN frame format


117/182


Type

2-byte Tag Protocol Identifier

(value: 81-00)

Tag Control Information (12 bit VLAN ID field,

3 bit priority field like IP TOS)

RecomputedCRC

Q f f

802.1 frame

802.1Q frame

Link Layer


118/182


L L y





5.5 Ethernet




Point to Point Data Link Control


119/182


one sender, one receiver, one link: easier thanbroadcast link:

no Media Access Control

no need for explicit MAC addressing

e.g., dialup link, ISDN line popular point-to-point DLC protocols:

PPP (point-to-point protocol)

HDLC: High level data link control (Data link

used to be considered high layer in protocolstack!

PPP Design Requirements [RFC 1557]


120/182


packet framing: encapsulation of network-layerdatagram in data link frame carry network layer data of any network layer

protocol (not just IP) at same time

ability to demultiplex upwards

bit transparency: must carry any bit pattern in thedata field

error detection (no correction)

connection liveness: detect, signal link failure tonetwork layer

network layer address negotiation: endpoint canlearn/configure each others network address

PPP non-requirements


121/182


q

no error correction/recovery no flow control

out of order delivery OK

no need to support multipoint links (e.g., polling)

Error recovery, flow control, data re-orderingall relegated to higher layers!

PPP Data Frame


122/182


Flag: delimiter (framing) Address: does nothing (only one option)

Control: does nothing; in the future possiblemultiple control fields

Protocol: upper layer protocol to which framedelivered (e.g., PPP-LCP, IP, IPCP, etc)

PPP Data Frame


123/182


info: upper layer data being carried check: cyclic redundancy check for error

detection

Byte Stuffing


124/182


y g data transparency requirement: data field must

be allowed to include flag pattern Q: is received data or flag?

Sender:adds (stuffs) extra < 01111110> byteafter each < 01111110> data byte

Receiver:

two 01111110 bytes in a row: discard first byte,continue data reception

single 01111110: flag byte

Byte Stuffing


125/182


y g

flag bytepatternin datato send

flag byte pattern plusstuffed byte intransmitted data

PPP Data Control Protocol


126/182


Before exchanging network-layer

data, data link peers must configure PPP link (max. frame

length, authentication)

learn/configure networklayer information

for IP: carry IP ControlProtocol (IPCP) msgs(protocol field: 8021) toconfigure/learn IP address

Link Layer


127/182


y


5.2 Error detectionand correction



5.5 Ethernet

5.6 Link-layer switches 5.7 PPP

Skip 5.8 and 5.9

5.8 Link virtualization: MPLS 5.9 A day in the life of a

web request


Virtualization of networks


128/182


Virtualization of resources: a powerful abstraction in

systems engineering: computing examples: virtual memory, virtual devices

Virtual machines: e.g., java

IBM VM OS from 1960s/70s

layering of abstractions: dont sweat the details ofthe lower layer, only deal with lower layers abstractly

The Internet: virtualizing networks


129/182


1974: multiple unconnectednets ARPAnet

data-over-cable networks

packet satellite network (Aloha)

packet radio network

differing in: addressing conventions packet formats

error recovery

routing

ARPAnet satellite net"A Protocol for Packet Network Intercommunication",

V. Cerf, R. Kahn, IEEE Transactions on Communications,

May, 1974, pp. 637-648.

The Internet: virtualizing networks


130/182


ARPAnet satellite net

gateway

Internetwork layer (IP): addressing: internetwork

appears as single, uniformentity, despite underlying localnetwork heterogeneity

network of networks

Gateway: embed internetwork packets in

local packet format or extractthem route (at internetwork level) to

next gateway

Cerf & Kahns Internetwork Architecture


131/182


What is virtualized? two layers of addressing: internetwork and local

network

new layer (IP) makes everything homogeneous atinternetwork layer

underlying local network technology cable

satellite

56K telephone modem today: ATM, MPLS

invisible at internetwork layer. Looks like a linklayer technology to IP!

ATM and MPLS


132/182


ATM, MPLS separate networks in their ownright different service models, addressing, routing

from Internet

viewed by Internet as logical link connectingIP routers just like dialup link is really part of separate

network (telephone network)

ATM, MPLS: of technical interest in theirown right

Asynchronous Transfer Mode: ATM


133/182


1990s/00 standard for high-speed (155Mbps to

622 Mbps and higher) Broadband IntegratedService Digital Networkarchitecture

Goal:integrated, end-end transport of carry voice,video, data

meeting timing/QoS requirements of voice, video(versus Internet best-effort model)

next generation telephony: technical roots intelephone world

packet-switching (fixed length packets, calledcells) using virtual circuits

Multiprotocol label switching (MPLS)


134/182


initial goal: speed up IP forwarding by using fixedlength label (instead of IP address) to doforwarding borrowing ideas from Virtual Circuit (VC) approach

but IP datagram still keeps IP address!

PPP or Ethernet

headerIP header remainder of link-layer frameMPLS header

label Exp S TTL

20 3 1 5

MPLS capable routers


135/182


a.k.a. label-switched router forwards packets to outgoing interface based

only on label value (dont inspect IP address) MPLS forwarding table distinct from IP forwarding

tables signaling protocol needed to set up forwarding RSVP-TE forwarding possible along paths that IP alone would

not allow (e.g., source-specific routing) !! use MPLS for traffic engineering

must co-exist with IP-only routers

MPLS forwarding tables


136/182


R1R2

D

R3R4R5

0

1

00

A

R6

in out out

label label dest interface

6 - A 0

in out out


10 6 A 1

12 9 D 0

in out out


10 A 0

12 D 0

1

in out out


8 6 A 0

0

8 A 1

Link Layer


137/182






5.5 Ethernet




Synthesis: a day in the life of a web request


138/182


journey down protocol stack complete! application, transport, network, link

putting-it-all-together: synthesis! goal:identify, review, understand protocols (at

all layers) involved in seemingly simple scenario:requesting www page

scenario:student attaches laptop to campusnetwork, requests/receives www.google.com

A day in the life: scenario


139/182


Comcast network68.80.0.0/13

Googles network

64.233.160.0/1964.233.169.105

web server

DNS server

school network

68.80.2.0/24

browser

web page

A day in the life connecting to the Internet

DHCP


140/182


connecting laptop needs toget its own IP address,

addr of first-hop router,addr of DNS server: useDHCP

router(runs DHCP)

DHCP

UDP

IP

Eth

Phy

DHCP

DHCP

DHCP

DHCP

DHCP

DHCP

UDPIP

Eth

Phy

DHCP

DHCP

DHCP

DHCPDHCP

DHCP request encapsulatedin UDP, encapsulated in IP,

encapsulated in 802.1Ethernet

Ethernet frame broadcast(dest: FFFFFFFFFFFF) on LAN,received at router running

DHCPserver Ethernet demuxedto IP

demuxed, UDP demuxed toDHCP

A day in the life connecting to the Internet

DHCP f lDHCP


141/182


DHCP server formulatesDHCP ACKcontainingclients IP address, IPaddress of first-hoprouter for client, name &IP address of DNS server

router(runs DHCP)

DHCP

UDP

IP

Eth

Phy

DHCP

DHCP

DHCP

DHCP

DHCP

UDPIP

Eth

Phy

DHCP

DHCP

DHCP

DHCP

DHCP

encapsulation at DHCPserver, frame forwarded(switch learning) throughLAN, demultiplexing atclient

Client now has IP address, knows name & addr of DNSserver, IP address of its first-hop router

DHCP client receives DHCPACK reply

A day in the life ARP (before DNS, before HTTP)

before sending HTTP requestDNS


142/182


before sending HTTPrequest,need IP address of www.google.com:

DNS

DNS

UDP

IP

Eth

Phy

DNS

DNS

DNS

DNS query created, encapsulatedin UDP, encapsulated in IP,encapsulated in Eth. In order tosend frame to router, need MACaddress of router interface: ARP

ARP querybroadcast, receivedby router, which replies withARP replygiving MAC addressof router interface

client now knows MAC addressof first hop router, so can nowsend frame containing DNSquery

ARP query

Eth

Phy

ARP

ARP

ARP reply

A day in the life using DNS

DNS

DNS server

DNS

UDP

IPDNS

DNS


143/182


DNS

UDP

IP

Eth

Phy

DNS

DNS

DNS

DNS

DNS

IP datagram containing DNSquery forwarded via LAN

switch from client to 1st hoprouter

IP datagram forwarded fromcampus network into comcastnetwork, routed (tables createdby RIP, OSPF, IS-ISand/orBGProuting protocols) to DNSserver

demuxed to DNS server DNS server replies to

client with IP address ofwww.google.com

Comcast network

68.80.0.0/13

Eth

Phy

DNS

DNS

A day in the life TCP connection carrying HTTP

HTTPHTTP


144/182


HTTP

TCP

IP

Eth

Phy

HTTP

to send HTTP request,

client first opens TCPsocketto web server TCP SYN segment(step 1

in 3-way handshake) inter-domain routedto webserver

TCP connection established!64.233.169.105web server

SYN

SYN

SYN

SYN

TCP

IP

Eth

Phy

SYN

SYN

SYN

SYNACK

SYNACK

SYNACK

SYNACK

SYNACK

SYNACK

SYNACK

web server responds withTCP SYNACK(step 2 in 3-way handshake)

A day in the life HTTP request/reply

HTTPHTTP

HTTP

web page finally (!!!)di l d


145/182


HTTP

TCP

IP

Eth

Phy

HTTP

HTTP requestsent intoTCP socket

IP datagram containingHTTP request routed towww.google.com

IP datagram containingHTTP reply routed back toclient

64.233.169.105

web server

HTTP

TCP

IP

Eth

Phy

web server responds withHTTP reply(containingweb page)

HTTP

HTTP

HTTPHTTP

HTTP

HTTP

HTTP

HTTP

HTTP

HTTP

HTTP

HTTP

HTTP

displayed

Chapter 5: Summary


146/182


principles behind data link layer services: error detection, correction sharing a broadcast channel: multiple access

link layer addressing

instantiation and implementation of various link

layer technologies Ethernet

switched LANS, VLANs

PPP

virtualized networks as a link layer: MPLS

synthesis: a day in the life of a web request

Chapter 5: lets take a breath


147/182


journey down protocol stack complete(except PHY)

solid understanding of networking principles,practice

.. could stop here . but lotsof interestingtopics! wireless

multimedia

security

network management

3.4 Principles of Reliable data transfer


148/182

TransportLayer3-148

important in app., transport, link layers top-10 list of important networking topics!

characteristics of unreliable channel will determinecomplexity of reliable data transfer protocol (rdt)

Network

Linklayer

Linklayer

Principles of Reliable data transfer


149/182

TransportLayer3-149

important in app., transport, link layers

top-10 list of important networking topics!


Network

Linklayer

Principles of Reliable data transfer


150/182

TransportLayer3-150

important in app., transport, link layers

top-10 list of important networking topics!


Network

Linklayer

Reliable data transfer: getting started


151/182

Transport Layer3-151

sendside receiveside

rdt_send():called from above,

(e.g., by app.). Passed data todeliver to receiver upper layer

udt_send():called by rdt,to transfer packet over

unreliable channel to receiver

rdt_rcv():called when packet

arrives on rcv-side of channel

deliver_data():called by

rdt to deliver data to upper

Reliable data transfer: getting started


152/182


Well:

incrementally develop sender, receiver sides ofreliable data transfer protocol (rdt)

consider only unidirectional data transfer but control info will flow on both directions!

use finite state machines (FSM) to specifysender, receiver

state1 state2

event causing state transitionactions taken on state transition

state:when in this

state next stateuniquely determined

by next event

eventactions

Rdt1.0: reliable transfer over a reliable channel


153/182


underlying channel perfectly reliable no bit errors no loss of packets

separate FSMs for sender, receiver: sender sends data into underlying channel

receiver read data from underlying channel

Wait forcall from

abovepacket = make_pkt(data)udt_send(packet)

rdt_send(data)

extract (packet,data)

deliver_data(data)

Wait forcall from

below

rdt_rcv(packet)

sender receiver

Rdt2.0: channel with bit errors


154/182


underlying channel may flip bits in packet checksum to detect bit errors

thequestion: how to recover from errors: acknowledgements (ACKs):receiver explicitly tells sender

that pkt received OK

negative acknowledgements (NAKs):receiver explicitlytells sender that pkt had errors

sender retransmits pkt on receipt of NAK

new mechanisms in rdt2.0 (beyond rdt1.0): error detection

receiver feedback: control msgs (ACK,NAK) rcvr->sender

How do humans recover from errorsduring conversation?

Rdt2.0: channel with bit errors


155/182


underlying channel may flip bits in packet checksum to detect bit errors

thequestion: how to recover from errors: acknowledgements (ACKs):receiver explicitly tells sender

that pkt received OK

negative acknowledgements (NAKs):receiver explicitlytells sender that pkt had errors

sender retransmits pkt on receipt of NAK

new mechanisms in rdt2.0 (beyond rdt1.0): error detection

receiver feedback: control msgs (ACK,NAK) rcvr->sender

rdt2.0: FSM specification

receiverrdt_send(data)


156/182


Wait for

call fromabove

sndpkt = make_pkt(data, checksum)

udt_send(sndpkt)

extract(rcvpkt,data) ,deliver_data(data)udt_send(ACK)

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)

rdt_rcv(rcvpkt) && isACK(rcvpkt)

udt_send(sndpkt)

rdt_rcv(rcvpkt) &&isNAK(rcvpkt)

udt_send(NAK)

rdt_rcv(rcvpkt) && corrupt(rcvpkt)

Wait for

ACK orNAK

Wait for call

from below

sender

receiver

L

rdt2.0: operation with no errors

rdt_send(data)


157/182


Wait for

call fromabove

snkpkt = make_pkt(data, checksum)

udt_send(sndpkt)

extract(rcvpkt,data),deliver_data(data)udt_send(ACK)



udt_send(sndpkt)

rdt_rcv(rcvpkt) &&isNAK(rcvpkt)

udt_send(NAK)

rdt_rcv(rcvpkt) && corrupt(rcvpkt)Wait forACK or

NAK

Wait forcall frombelow

L

rdt2.0: error scenariordt_send(data)


158/182


Wait for

call fromabove

snkpkt = make_pkt(data, checksum)udt_send(sndpkt)

extract(rcvpkt,data),deliver_data(data)udt_send(ACK)



udt_send(sndpkt)

rdt_rcv(rcvpkt) && isNAK(rcvpkt)

udt_send(NAK)

rdt_rcv(rcvpkt) && corrupt(rcvpkt)Wait for

ACK orNAK

Wait for

call frombelow

L

rdt2.0 has a fatal flaw!


159/182


What happens ifACK/NAK corrupted? sender doesnt know what

happened at receiver!

cant just retransmit:

possible duplicate

Handling duplicates: sender retransmits current

pkt if ACK/NAK garbled

sender adds sequencenumberto each pkt

receiver discards (doesntdeliver up) duplicate pkt

Sender sends one packet,then waits for receiverresponse

stop and wait

rdt2.1: sender, handles garbled ACK/NAKs

rdt_send(data)


160/182


Wait forcall0fromabove

sndpkt = make_pkt(0, data, checksum)

udt_send(sndpkt)

Wait forACK orNAK 0

udt_send(sndpkt)

rdt_rcv(rcvpkt) &&( corrupt(rcvpkt) ||isNAK(rcvpkt) )

sndpkt = make_pkt(1, data, checksum)udt_send(sndpkt)

rdt_send(data)

rdt_rcv(rcvpkt)&& notcorrupt(rcvpkt)&& isACK(rcvpkt)

udt_send(sndpkt)

rdt_rcv(rcvpkt) &&

( corrupt(rcvpkt) ||isNAK(rcvpkt) )

rdt_rcv(rcvpkt)&& notcorrupt(rcvpkt)

&& isACK(rcvpkt)

Wait forcall1 fromabove

Wait for ACKor NAK 1

L

L

rdt2.1: receiver, handles garbled ACK/NAKs

d ( k ) ( k )


161/182


Wait for0 frombelow

sndpkt = make_pkt(NAK, chksum)udt_send(sndpkt)

rdt_rcv(rcvpkt) &&not corrupt(rcvpkt) &&has_seq0(rcvpkt)


&& has_seq1(rcvpkt)extract(rcvpkt,data)deliver_data(data)sndpkt = make_pkt(ACK, chksum)udt_send(sndpkt)

Wait for1 frombelow

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt)&& has_seq0(rcvpkt)

extract(rcvpkt,data)deliver_data(data)sndpkt = make_pkt(ACK, chksum)udt_send(sndpkt)

rdt_rcv(rcvpkt) && (corrupt(rcvpkt)

sndpkt = make_pkt(ACK, chksum)udt_send(sndpkt)

rdt_rcv(rcvpkt) &&not corrupt(rcvpkt) &&has_seq1(rcvpkt)

rdt_rcv(rcvpkt) && (corrupt(rcvpkt)

sndpkt = make_pkt(ACK, chksum)udt_send(sndpkt)

sndpkt = make_pkt(NAK, chksum)udt_send(sndpkt)

rdt2.1: discussion


162/182


Sender: seq # added to pkt

two seq. #s (0,1) willsuffice. Why?

must check if receivedACK/NAK corrupted

twice as many states state must remember

whether current pkthas 0 or 1 seq. #

Receiver: must check if received

packet is duplicate state indicates whether

0 or 1 is expected pktseq #

note: receiver can notknow if its lastACK/NAK received OK

at sender

rdt2.2: a NAK-free protocol


163/182


same functionality as rdt2.1, using ACKs only instead of NAK, receiver sends ACK for last pkt

received OK receiver must explicitlyinclude seq # of pkt being ACKed

duplicate ACK at sender results in same action asNAK: retransmit current pkt

rdt2.2: sender, receiver fragments

rdt send(data)


164/182


Wait for call0 fromabove

sndpkt = make_pkt(0, data, checksum)

udt_send(sndpkt)

rdt_send(data)

udt_send(sndpkt)

rdt_rcv(rcvpkt) &&

(corrupt(rcvpkt) ||isACK(rcvpkt,1))

rdt_rcv(rcvpkt)

&& notcorrupt(rcvpkt)&& isACK(rcvpkt,0)

Wait forACK0

sender FSMfragment

rdt_rcv(rcvpkt) && notcorrupt(rcvpkt) && has_seq1(rcvpkt)

extract(rcvpkt,data)deliver_data(data)sndpkt = make_pkt(ACK1, chksum)udt_send(sndpkt)

Wait for0 frombelow

rdt_rcv(rcvpkt) &&(corrupt(rcvpkt) ||

has_seq1(rcvpkt))

udt_send(sndpkt)

receiver FSMfragment

L

rdt3.0: channels with errors andloss


165/182


New assumption:underlying channel canalso lose packets (dataor ACKs) checksum, seq. #, ACKs,

retransmissions will beof help, but not enough

Approach: sender waitsreasonable amount oftime for ACK

retransmits if no ACKreceived in this time

if pkt (or ACK) just delayed(not lost):

retransmission will beduplicate, but use of seq.#s already handles this

receiver must specify seq# of pkt being ACKed

requires countdown timer

rdt3.0 senderrdt_send(data)

rdt rcv(rcvpkt) &&


166/182


sndpkt = make_pkt(0, data, checksum)udt_send(sndpkt)

start_timer

WaitforACK0

dt_ c ( c p t) &&( corrupt(rcvpkt) ||isACK(rcvpkt,1) )

Wait forcall 1 fromabove

sndpkt = make_pkt(1, data, checksum)udt_send(sndpkt)start_timer

rdt_send(data)

rdt_rcv(rcvpkt)

&& notcorrupt(rcvpkt)&& isACK(rcvpkt,0)

rdt_rcv(rcvpkt) &&( corrupt(rcvpkt) ||isACK(rcvpkt,0) )

rdt_rcv(rcvpkt)&& notcorrupt(rcvpkt)

&& isACK(rcvpkt,1)

stop_timer

stop_timer

udt_send(sndpkt)start_timer

timeout

udt_send(sndpkt)start_timer

timeout

rdt_rcv(rcvpkt)

Wait forcall 0fromabove

WaitforACK1

L

rdt_rcv(rcvpkt)

L

L

L

rdt3.0 in action


167/182

TransportLayer3-167

rdt3.0 in action


168/182

TransportLayer3-168

Performance of rdt3.0


169/182

TransportLayer3-169

rdt3.0 works, but performance stinks ex: 1 Gbps link, 15 ms prop. delay, 8000 bit packet:

Usender:utilizationfraction of time sender busy sending

Usender

=.008

30.008= 0.00027

L / R

RTT + L / R=

if RTT=30 msec, 1KB pkt every 30 msec -> 33kB/sec thruputover 1 Gbps link

network protocol limits use of physical resources!

dsmicrosecon8bps10

bits80009

R

Ldtrans

rdt3.0: stop-and-wait operation


170/182

TransportLayer3-170

first packet bit transmitted, t =0

sender receive

r

RTT

last packet bit transmitted, t = L/ R

first packet bit arriveslast packet bit arrives, send

ACK

ACK arrives, send nextpacket, t = RTT + L / R

Usender

=.008

30.008= 0.00027

L / R

RTT + L / R=

Pipelined protocols

i li i d ll lti l i fli ht t t


171/182

TransportLayer3-171

pipelining:sender allows multiple, in-flight, yet-to-

be-acknowledged pkts range of sequence numbers must be increased

buffering at sender and/or receiver

two generic forms of pipelined protocols:go-Back-N,selective repeat

Pipelining: increased utilization

sender receiver


172/182

TransportLayer3-172

first packet bit transmitted, t

= 0

sender

RTT

last bit transmitted, t = L /R

first packet bit arriveslast packet bit arrives, send ACK

ACK arrives, send nextpacket, t = RTT + L / R

last bit of 2ndpkt arrives, send ACK

last bit of 3rdpkt arrives, send ACK

Usender

=.024

30.008= 0.0008

3 * L / R

RTT + L / R=

Increase utilizationby a factor of 3!

Pipelined Protocols


173/182

TransportLayer3-173

Go-back-N: big picture: sender can have up to N

unacked packets inpipeline

rcvr only sends

cumulative acks doesnt ack packet if

theres a gap

sender has timer foroldest unacked packet if timer expires,

retransmit allunackedpackets

Selective Repeat: big pic sender can have up to N

unacked packets inpipeline

rcvr sends individual ack

for each packet

sender maintains timerfor each unacked packet

when a timer expires,retransmit onlyunackedpacket

Go-Back-NS d :


174/182

TransportLayer3-174

Sender:

k-bit seq # in pkt header window of up to N, consecutive unacked pkts allowed

ACK(n): ACKs all pkts up to, including seq # n -cumulative ACK

may receive duplicate ACKs (see receiver) timer for each in-flight pkt

timeout(n): retransmit pkt n and all higher seq # pkts in window

GBN: sender extended FSMrdt_send(data)


175/182

TransportLayer3-175

Waitstart_timerudt_send(sndpkt[base])udt_send(sndpkt[base+1])udt_send(sndpkt[nextseqnum-1])

timeout

if (nextseqnum < base+N) {sndpkt[nextseqnum] = make_pkt(nextseqnum,data,chksum)

udt_send(sndpkt[nextseqnum])if (base == nextseqnum) start_timernextseqnum++

}elserefuse_data(data)

base = getacknum(rcvpkt)+1If (base == nextseqnum) stop_timerelse start_timer


base=1

nextseqnum=1

rdt_rcv(rcvpkt)&& corrupt(rcvpkt)

L

GBN: receiver extended FSMdefault


176/182

TransportLayer3-176

ACK-only: always send ACK for correctly-received pktwith highest in-orderseq # may generate duplicate ACKs need only remember expectedseqnum

out-of-order pkt: discard (dont buffer) -> no receiver buffering!

Re-ACK pkt with highest in-order seq #

Wait

udt_send(sndpkt)rdt_rcv(rcvpkt)

&& notcurrupt(rcvpkt)&& hasseqnum(rcvpkt,expectedseqnum)

extract(rcvpkt,data)deliver_data(data)sndpkt = make_pkt(expectedseqnum,ACK,chksum)udt_send(sndpkt)expectedseqnum++

expectedseqnum=1sndpkt =make_pkt(expectedseqnum,ACK,chksum)

L

GBN inaction


177/182

TransportLayer3-177

Selective Repeat


178/182

TransportLayer3-178

receiver individuallyacknowledges all correctlyreceived pkts buffers pkts, as needed, for eventual in-order delivery

to upper layer

sender only resends pkts for which ACK not

received sender timer for each unACKed pkt

sender window N consecutive seq #s

again limits seq #s of sent, unACKed pkts

Selective repeat: sender, receiver windows


179/182

TransportLayer3-179

Selective repeat

sender receiver


180/182

TransportLayer3-180

data from above : if next available seq # in

window, send pkt

timeout(n): resend pkt n, restart timer

ACK(n) in [sendbase,sendbase+N]: mark pkt n as received

if n smallest unACKed pkt,advance window base to

next unACKed seq #

sender

pkt n in[rcvbase, rcvbase+N-1] send ACK(n)

out-of-order: buffer

in-order: deliver (alsodeliver buffered, in-order

pkts), advance window tonext not-yet-received pkt

pkt n in[rcvbase-N,rcvbase-1] ACK(n)

otherwise: ignore

Selective repeat in action


181/182

TransportLayer3-181

Selective repeat:dilemma


182/182

Example: seq #s: 0, 1, 2, 3

window size=3

receiver sees no

difference in twoscenarios!

incorrectly passesduplicate data as newin (a)

chapter5(link) ece358slides

Documents